1 / 49

MINC meeting 2003 Registration techniques issues

MINC meeting 2003 Registration techniques issues. D. Louis Collins <louis@bic.mni.mcgill.ca>. Outline. Introduction to registration definitions motivation Stereotaxic Space Registration similarity measures transform types optimization procedures Methods Talairach, SPM, AIR, MRITOTAL

kalare
Download Presentation

MINC meeting 2003 Registration techniques issues

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. MINC meeting 2003Registration techniques issues D. Louis Collins<louis@bic.mni.mcgill.ca>

  2. Outline • Introduction to registration • definitions • motivation • Stereotaxic Space • Registration • similarity measures • transform types • optimization procedures • Methods • Talairach, SPM, AIR, MRITOTAL • Applications

  3. Registration • Registration is the process of alignment of medical imaging data (usually for the purpose of comparison). • Intra-subject: between data volumes from the same subject • Inter-subject: between data volumes from different subjects

  4. Motivation / Uses • image guided surgery • analysis of functional images • characterization of normal and abnormal anatomical variability • detection of change in disease state over time • visualization of multimodality data • modeling anatomy in the process of segmentation • atlas guidance for anatomical interpretation VIPER T Peters, K Finnis, D. Gobbi, Y Starreveld - RRI

  5. Motivation / Uses • image guided surgery • analysis of functional images • characterization of normal and abnormal anatomical variability • detection of change in disease state over time • visualization of multimodality data • modeling anatomy in the process of segmentation • atlas guidance for anatomical interpretation O. Rousset, A Evans - MNI

  6. Motivation / Uses • image guided surgery • analysis of functional images • characterization of normal and abnormal anatomical variability • detection of change in disease state over time • visualization of multimodality data • modeling anatomy in the process of segmentation • atlas guidance for anatomical interpretation L Collins (94) - MNI

  7. Motivation / Uses • image guided surgery • analysis of functional images • characterization of normal and abnormal anatomical variability • detection of change in disease state over time • visualization of multimodality data • modeling anatomy in the process of segmentation • atlas guidance for anatomical interpretation S Smith, P Matthews - Oxford

  8. Motivation / Uses • image guided surgery • analysis of functional images • characterization of normal and abnormal anatomical variability • detection of change in disease state over time • visualization of multimodality data • modeling anatomy in the process of segmentation • atlas guidance for anatomical interpretation register program - MNI

  9. Motivation / Uses • image guided surgery • analysis of functional images • characterization of normal and abnormal anatomical variability • detection of change in disease state over time • visualization of multimodality data • modeling anatomy in the process of segmentation • atlas guidance for anatomical interpretation W. Nowinski - KRDL

  10. Motivation / Uses • image guided surgery • analysis of functional images • characterization of normal and abnormal anatomical variability • detection of change in disease state over time • visualization of multimodality data • modeling anatomy in the process of segmentation • atlas guidance for anatomical interpretation Talairach Atlas overlaid on MRI

  11. Inter-subject registration requires a well defined target space.

  12. Stereotaxic Space J. Talairach and P. Tournoux, Co-planar stereotactic atlas of the human brain: 3-Dimensional proportional system: an approach to cerebral imaging, Stuttgart, Georg Thieme Verlag, 1988 • based on anatomical landmarks (anterior and posterior commissures) • originally used to guide blind stereotaxic neurosurgical procedures (thalamotomy, pallidotomy) • now used by NeuroScientific community for interpretation and comparison of results

  13. AC-PC line posterior commissure AC-PC line anterior commissure VAC

  14. Stereotaxic Space J Talairach & P Tournoux, Co-planar stereotaxic atlas of the human brain, Georg Thieme, 1988

  15. Stereotaxic Space

  16. Talairach Atlas • is derived from an unrepresentative single 60-yr old female cadaver brain (when most functional activation studies are done on young living subjects!) • ignores left-right hemispheric differences • has variable slice separation, up to 4mm • while it contains transverse, coronal and sagittal slices, it is not contiguous in 3D Drawbacks for functional imaging:

  17. Stereotaxic Space Advantages for functional imaging: • Provides a conceptual framework for the completely automated, 3D analysis across subjects. • Facilitate intra/inter-subject comparisons across • time points, subjects, groups, sites • Extrapolate findings to the population as a whole • Increase activation signal above that obtained from single subject • Increase number of possible degrees of freedom allowed in statistical model • Enable reporting of activations as co-ordinates within a known standard space • e.g. the space described by Talairach & Tournoux

  18. Stereotaxic Space • Allows the use of spatial masks for post-processing (anatomically driven hypothesis testing) • allows the use of spatial priors (classification) • allows the use of anatomical models (segmentation) • provides a framework for statistical analysis with well-established random field models • Allows the rapid re-analysis using different criteria Advantages (continued):

  19. Registration Requirements: 1- similarity measure how to define the match? what is the goal? 2- well defined transformationhow to define the mapping? 3- method to find transformationhow to find the mapping given the similarity constraint?

  20. Similarity Measures 0D - points 1D - lines 2D - surfaces 3D - volumes nD - data over time • Extrinsicframes, moulds, masks, markers • Intrinsicanatomical landmarks • Non-image dataacquisition based Review: P. van den Elsen, “Medical Image registration: a review with classification”, IEEE Eng in Med & Biol, 1993 12(1):26-39

  21. Point Similarity Measures • Requires identification of homologous landmark points • Based on minimization of distance between points Error T T found by SVD or Procrustes Number of points

  22. Line Similarity Measures • Based on distance between homologous lines • Used for intra-subject registration • Difficult to use in inter-subject registration due to (lack of) homology G. Subsol, INRIA

  23. S B S A x B i x A i c A å = x T x min [ , ( )] D d A B i i S i T Surface Similarity Measures • Based on distance between surfaces • need to ensure that the same anatomical surface is extracted from both data sets "Head-and-hat" 1. Segment slices to get S contours. A Compute centroid of S : c . A A 2. For each x , find inter section x B i A i along path to c . A 3. Pelizzari CA, Chen GTY, Spelbring DR, Weichselbaum RR, Chen C-T. Accurate three-dimensional registration of CT, PET, and/or MR images of the brain. J Comput Assist Tomogr 1989;13(1):20-26

  24. Surface based registration Surface model Surface data Surface data matched to model Local geometry constraints Randy Ellis, Queens U. A Johnson, Robotic Inst., CMU

  25. Volume Similarity Measures • The pixel/voxel intensities are used directly • to compute the similarity measure • Intra-modality (same modality) • similar contrast • similar resolution • similar sampling (pixel/voxel size) • similar structures have similar intensities • Inter-modality • different contrast • different resolution • different sampling (pixel/voxel size) • different structures may have similar intensities, and similar structures may have the same intensity

  26. Volume Similarity Measures t v d INTRA-MODALITY • Absolute or squared difference • Hoh93, Lange93, Christensen95, Hajnal95, Kruggel95 • Stochastic Sign Change (SSC),Deterministic Sign Change (DSC) • Venot83, Minoshima92, Hua93, Hoh93 • Cross Correlation • Junck90, van den Elsen93, Hill93,Collins94, Lemieux94, Studholme95 • Fourier Domain Correlation • de Castro87, Leclerc87, Chen93,Lehmann96 • Optic Flow Field • Barber95, Meunier96 • Very simple (fast) to compute • Must have similar intensities • Unbounded maximum value

  27. Volume Similarity Measures t v d INTRA-MODALITY • Absolute or squared difference • Hoh93, Lange93, Christensen95, Hajnal95, Kruggel95 • Stochastic Sign Change (SSC),Deterministic Sign Change (DSC) • Venot83, Minoshima92, Hua93, Hoh93 • Cross Correlation • Junck90, van den Elsen93, Hill93,Collins94, Lemieux94, Studholme95 • Fourier Domain Correlation • de Castro87, Leclerc87, Chen93,Lehmann96 • Optic Flow Field • Barber95, Meunier96 • Very simple (fast) to compute • Must have similar intensities • Unbounded maximum value • Can add artificial noise if needed

  28. p Volume Similarity Measures t v INTRA-MODALITY • Absolute or squared difference • Hoh93, Lange93, Christensen95, Hajnal95, Kruggel95 • Stochastic Sign Change (SSC),Deterministic Sign Change (DSC) • Venot83, Minoshima92, Hua93, Hoh93 • Cross Correlation • Junck90, van den Elsen93, Hill93,Collins94, Lemieux94, Studholme95 • Fourier Domain Correlation • de Castro87, Leclerc87, Chen93,Lehmann96 • Optic Flow Field • Barber95, Meunier96 * 1.0 • Must have linear relation between intensities • Bounded value [0..1]

  29. Volume Similarity Measures INTER-MODALITY • Variance of Ratios • Woods92,93, Hill93, Zuo96 • Min. variance of ratios insegments • Cox94, Ardekani95 • Mutual Information/ Entropy • Collignon93, Studholme94 • Correlation Ratio • Roche98

  30. Volume Similarity Measures INTER-MODALITY • Variance of Ratios • Woods92,93, Hill93, Zuo96 • Min. variance of ratios insegments • Cox94, Ardekani95 • Mutual Information/ Entropy • Collignon93, Studholme94 • Correlation Ratio • Roche98 Where: - marginal probability distributions - joint probability distribution If statistically independent If related by 1:1 mapping T().

  31. Transformation Types Linear rigid body: 3 rotations, 3 translations Procrustes: 3 rotations, 3 translations, 1 scale affine: 3 rotations, 3 translations, 3 scale, 3 skew Piecewise Linear Talairach: 12 regions defined by 2 points + 6 scales Nonlinear polynomial: f(x) = ax^3 + bx^2 + cx + d basis functions: cosine, Fourier, wavelet physical model: elastic, fluid with dense deformation field

  32. mni_autoreg • Volumetric registration with minctracc • Linear • lsq6 (rigid body) • lsq7 (rigid + isotropic scale) • lsq9 (rigid + 3 scales) • Lsq12 (full affine) • Non-linear • Deformation field

  33. mni_autoreg: mritoself mritoself scan1.mnc scan2.mnc t1-2.xfm -veryclose same session -close simplex 3 -far same scanner, diff sessions -xcorr, -vr, -mi (default) -lsq6,-lsq7,-lsq9 -mask

  34. mni_autoreg: mritoself mritoself scan1.mnc scan2.mnc t1-2.xfm mincresample scan1.mnc scan1-like2.mnc \-transformation t1-2.xfm \-like scan2.mnc

  35. Stereotaxic Registration methods • Talairach Talairach and Tournoux • mritotal Collins • SPM Friston, Ashburner • FLIRT,FSL Jenkinson, Smith

  36. identify AC/PC on mid-sagittal define vertical, lateral and anterior-posterior extents define 12 piecewise linear transformations: left / right above / below AC-PC anterior-AC / AC-PC / PC-posterior Talairach superior right posterior anterior left inferior

  37. mritotal • Principal axis transformation • correlation of 16mm fwhm blurred data • correlation of 8mm fwhm blurred data • correlation of 8mm gradient magnitude data PAT http://www.bic.mni.mcgill.ca/software/mni_autoreg/ Collins et al, JCAT 1994

  38. mni_autoreg: mritotal mritotal scan1.mnc t_stx.xfm -crops, blurs -transformation -model mincresample scan1.mnc scan_stx.mnc \-transformation t_stx.xfm \-like stx_target.mnc

  39. FLIRT • Correlation ratio • Multi-resolution procedure • Powell’s search for optimmization Jenkinson, M. and Smith, S. (2001a). A global optimisation method for robust affine registration of brain images. Medical Image Analysis, 5(2):143-156

  40. SPM: Statistical Parametric Mapping Spatially normalised Spatial Normalisation Original image Determine the spatial transformation that minimises the sum of squared difference between an image and a linear combination of one or more templates. Begins with an affine registration to match the size and position of the image. Followed by a global non-linear warping to match the overall brain shape. Uses a Bayesian framework to simultaneously maximise the smoothness of the warps. Spatial Normalisation Template image J. Ashburner, FIL, London

  41. T1 Transm T2 T1 305 T2 PD SS PD PET EPI Template Images “Canonical” images A wider range of different contrasts can be normalised by registering to a linear combination of template images. Spatial normalisation can be weighted so that out of brain voxels do not influence the result. Similar weighting masks can be used for normalising lesioned brains. J. Ashburner, FIL, London

  42. Canonical Images • SPM • SPM96: average of 12 manually transformed vols • SPM97: blurred colin27, mni305 if downloaded • SPM99: mni305; colin27 option • SPM2b11RC: icbm152 • mritotal • mni305 • icbm152 • Flirt • mni305

  43. Y=-30 X=10 Z=-10 Y=0 X=20 Z=0 Z=20 Y=20 X=50 Examples: MNI305 average brain A.C. Evans et al, 1992

  44. Examples: ICBM152 averages Average T1 Average PD Average T2

  45. Canonical targets colin27 mni305 icbm152 child175 www.bic.mni.mcgill.ca/icbmview

  46. Things to take home • Mapping depends on • Similarity function • Target model • Optimization function/strategy • Use a standard model!

  47. fin

  48. Comparison Preliminary results from consistency study reveals differences in robustness In each graph the average rms error (in mm) is plotted over a set of initially rotated image volumes Steve Smith, FMRIB, Oxford

More Related